Download - Scholarworks @ Morehead State

IT-CLEFTS AND THEIR PROCESSING
__________________________
A Thesis
Presented to
the Faculty of the Caudill College of Arts, Humanities, and Social Sciences
Morehead State University
_________________________
In Partial Fulfillment
of the Requirements for the Degree
Master of Arts
_________________________
by
Elisabeth S. Johnson
July 21, 2015
ProQuest Number: 1602534
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
ProQuest 1602534
Published by ProQuest LLC (2015). Copyright of the Dissertation is held by the Author.
All rights reserved.
This work is protected against unauthorized copying under Title 17, United States Code
Microform Edition © ProQuest LLC.
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
Accepted by the faculty of the Caudill College of Arts, Humanities, and Social Sciences,
Morehead State University, in partial fulfillment of the requirements for the Master of
Arts degree.
____________________________
Katy Carlson
Director of Thesis
Master’s Committee:
________________________________, Chair
Katy Carlson
_________________________________
Annie Adams
_________________________________
Robert D. Royar
________________________
Date
IT-CLEFTS AND THEIR PROCESSING
Elisabeth S. Johnson
Morehead State University, 2015
Director of Thesis: __________________________________________________
Katy Carlson
This thesis is a study of the it-cleft structure, as in sentences like It was Jim who
ran the marathon. This sentence structure contains a cleft pronoun (It), a copular verb
(was), a cleft phrase (Jim), and a cleft clause (who ran the marathon). The thesis
investigates how people linguistically process this type of sentence and includes
research on subject-extracted cleft clauses versus object-extracted ones (e.g., It was Jim
whom the judges liked best.). Sentences also vary the use of nominative personal
pronouns (e.g., she/we) versus accusative personal pronouns (e.g., her/us) and
nominative versus accusative relative pronouns (who vs. whom).
Syntactically, it-cleft sentences can be analyzed in several ways. The syntactic
theories covered in this thesis are the extraposition approach, the expletive theory, and
the it-as-subject analysis. The extraposition approach maintains that the cleft clause is
connected with the cleft pronoun, not the cleft phrase. The expletive theory states that
the cleft pronoun is not present in the initial syntactic structure; its purpose is to place
emphasis on the cleft phrase. The it-as-subject analysis is distinguished by the linking
of the cleft phrase with the cleft clause, a connection that does not exist in either the
extraposition or the expletive theory. Even though the syntactic literature has not
settled on one theory of clefts, common structural features can be identified that affect
their processing.
The thesis also contains an overview of the processing of relative clauses
(because they are similar to cleft clauses) and cleft clauses. It summarizes research on
related pronoun issues in sentences, including a supposed wane in the use of whom in
relative or cleft clauses and the increasing acceptance of accusative personal pronouns
where prescriptive rules determine that nominative ones should be used.
The experiment conducted for this thesis was a combination of a sentence-rating
study and a fill-in study. The rating questionnaire contained subject and object clefts
varying the case (nominative or accusative) of both the personal and relative pronouns.
The fill-in portion provided participants with four non-cleft sentences and asked people
to complete it-cleft beginnings for each of those four sentences.
The results show that people prefer accusative personal pronouns (e.g., me, them)
instead of the prescriptively required nominative personal pronouns (e.g., I, they) in itcleft sentences. The experiment also supports earlier research that subject-extracted
clefts are more easily processed than object-extracted clefts. The misuse of whom in the
fill-in sentences was not as prevalent as expected, though the average rating for
sentences containing whom was much higher for those subjects who did misuse whom
than the rating of those who used whom correctly in the fill-in portion of the experiment.
The results also show that there are still a good number of students who don’t
understand the rules governing the use of who and whom.
The process of researching and writing this thesis, and conducting the
experiment for it, gave me insight into how students understand it-cleft sentences
specifically, and, generally, how they understand the differences between nominative
and accusative pronouns and relative pronouns and where each should be used. It
showed me where teachers can be flexible about certain prescriptive rules and where,
for the purposes of formal writing, at least, some rules should still be enforced even
though they may be difficult to understand and to put into use.
Accepted by:
________________________________, Chair
Katy Carlson
_________________________________
Annie Adams
_________________________________
Robert D. Royar
DEDICATION
This thesis is dedicated to my husband, Charles, who helped me through more
than three years of going to school at the same time as working full time. He cooked
meals when we’d both had long days at work so I could do my homework, he dried my
tears when I was stressing out over that same homework, and he was generally my rock
as I completed my degree. I could not have done this without his support.
ACKNOWLEDGEMENTS
I would like to thank my thesis advisor, Dr. Katy Carlson, for her steadfast
support and assistance throughout the research and writing of my thesis. It was her
two linguistics courses that set me on the path for this journey. She appreciated my
love of the English language and the intricacies of grammar rules, and was instrumental
in turning that fascination into a viable idea. She prodded me through my
procrastination, encouraged me, and gave me confidence that I could indeed complete
this momentous task. I would also like to thank the other members of my thesis
committee, Dr. Annie Adams and Dr. Robert Royar, for agreeing to take the time to
serve on my committee and, with Dr. Carlson, for making my thesis defense fun. It was
one of the most enjoyable things I’ve experienced.
Last, but certainly not least, I’d like to thank two friends and former colleagues in
particular: Andrew DiNicola and Hayley Scheeser. They read drafts of my prospectus
and thesis, served as sounding boards when I got writer’s block, and were sources of
great encouragement.
TABLE OF CONTENTS
Chapter I:
Definition and Purpose
1
Chapter II:
Syntax
4
Chapter III: Processing
16
Chapter IV: Experiment
27
Chapter V:
44
Discussion
References
51
Appendix A: Experiment Sentences
53
Appendix B Minor Patterns
57
LIST OF TABLES
Table 1:
Average ratings of all participants
32
Table 2:
Actual sentence responses through critical verb
37
Table 3:
Breakdown of cleft clause relative pronoun responses
39
Table 4:
Average ratings of participants misusing whom in fill-in clefts
41
Table 5:
Average ratings of participants NOT misusing whom in fill-in clefts
41
Table 6:
Average ratings of participants using only that in fill-in clefts
42
Table 7:
Average rating of participants changing object to subject cleft
43
IT-CLEFTS AND THEIR PROCESSING
Chapter I:
Definition and Purpose
It-clefts, generally, are sentences, as in (1), that contain a cleft pronoun, a copular
verb [form of to be], a cleft phrase, and a cleft clause:
(1)
It
was
Shanna
who fell.
cleft pronoun copular verb cleft phrase cleft clause
The purpose of this thesis project is to study how people process/understand a
specific type of it-cleft sentence. These sentences will contain either subject cleft clauses
or object cleft clauses. In a subject cleft such as (2), the apparent relative pronoun who
replaces the subject of the verb completed in the cleft clause, but in an object cleft such as
(3), the pronoun whom replaces the object of the verb thanked in the cleft clause.
(2)
It was I who completed the assignment.
[cleft clause refers to subject I]
(3)
It was we whom the mayor thanked.
[cleft clause refers to object we]
As background to the processing study, I will be reviewing several syntactic
proposals for the cleft structure; how to treat this structure is an interesting unsolved
problem within syntax. While the theories differ, there are, however, some common
ideas that relate to the processing of the clefts. My review is based on the generative
grammar theory of Principles and Parameters as covered in Andrew Carnie’s Syntax: A
Generative Introduction.
Additionally, the study will explore how prescriptive rules dictate the use of
pronouns as compared to how people actually use those pronouns. For instance, both
(2) and (3) follow the grammar rule of having the cleft phrase (I/we) be a nominative
IT-CLEFTS AND THEIR PROCESSING
2
personal pronoun to match the fact that, prescriptively, the it is considered a subject
pronoun. More frequently, however, people tend to readily accept accusative pronouns
in the cleft phrase position, as in (4) and (5).
(4)
It was me who completed the assignment.
(5)
It was us whom the mayor thanked.
There is also a prescriptive rule about the relative pronouns who (nominative)
and whom (accusative) and when they should be used in sentences. The apparent
decline in the use of the pronoun whom, as discussed in Aarts (1994) and demonstrated
by the acceptance of sentences such as (6), raises the question of how such a decline
might affect how people rate sentences containing whom, as well as how they accept
sentences that do and don’t use whom in a manner consistent with prescriptive
grammar.
(6)
It was we/us who the mayor thanked.
For example, might people reading or hearing sentences like (3) or (5) think those
sentences are unnatural or ungrammatical, perhaps because they are used to using who
(a nominative pronoun) even in sentences where an accusative pronoun (in this case,
whom) should be used? Or might they misuse whom (for example, use a sentence like *It
was he whom bought the new car) because they are attempting to sound more educated?
Do people gravitate towards the use of that, which in formal rules of grammar is
supposed to be used only for groups or things, not people, to avoid the who/whom
conundrum? This thesis hopes to address these types of questions through a rating
study and an associated fill-in mini-experiment.
IT-CLEFTS AND THEIR PROCESSING
3
As a teacher of English grammar, in investigating these questions, I would like to
gain knowledge of how students, especially, deal with the it-cleft sentence and how
they process both the cleft phrase personal pronouns and the relative pronouns present
in these clefts. Do students understand how prescriptive rules dictate that these
pronouns (both nominative and accusative cases) should be used? Are they able to
actually use these pronouns correctly? I hope that the information I get from this study
will assist me in teaching these rules more effectively to my students.
IT-CLEFTS AND THEIR PROCESSING
Chapter II:
4
Syntax
It-cleft sentences have a complex structure, and there are several different
theories that attempt to explain them. None of these theories, however, is widely
accepted by the syntactic community. All of the syntactic theories about it-clefts assume
two important properties of clefts and how they relate to meaning:
(a) The it–cleft is a focusing construction: the focal position, which occurs after
the copular verb, gives the main information of the sentence and is frequently
stressed. In other words, the it-cleft has the foregrounded information in the
postcopular position and remaining semantic information is in a sentencefinal clause (the cleft clause) (Patten, 2012).
(b) it-clefts exhibit exhaustiveness: in (7) an assumption is made that Shanna is
the only one who fell. In (8), the negation lets us know that Shanna didn’t fall,
but we have to assume that someone else did (Patten, 2012).
(7) It was Shanna who fell.
(8) It wasn’t Shanna who fell.
An overview of three approaches to the structure of it-clefts follows: the
extraposition, the expletive, and the it-as-subject theories. While there are similarities
among the approaches, the analysis of every element of a cleft sentence is controversial
and varies in each of these theories.
The extraposition approach: This approach relates it-clefts to specificational
copular sentences (9), instead of to simple non-cleft sentences (10).
IT-CLEFTS AND THEIR PROCESSING
5
(9) The one that complained was Frank.
(10) Frank complained.
(Patten, 2012 p. 2)
In an early paper, Jespersen (1927, as cited in Patten), was a proponent of the
extraposition approach. He developed a transposition account of it-clefts in which he
considered that the cleft clause (e.g., who fell in (7)) is a restrictive relative one, which
modifies the cleft pronoun (it) instead of the cleft phrase. According to this analysis, It is
John that Mary saw really means that the relative clause that Mary saw belongs to it rather
than to what follows it (Patten, 2012 p. 8). The tree in (11) is based on the work of more
current theorists rather than directly on Jespersen’s ideas.
(11) Extraposition Approach (Hartmann and Veenstra, 2013 p. 9; Patten, 2012 p. 110)
IT-CLEFTS AND THEIR PROCESSING
6
IT-CLEFTS AND THEIR PROCESSING
7
Unlike in the expletive approach discussed below (15), that Mary saw is not an
argument of the copular verb, though John is an argument of the verb be in both
approaches. The cleft clause forms a definite NP, which is a discontinuous constituent
with the cleft pronoun (see D-structure and Jespersen’s transposition account above); it
is taken to be referential (Hartmann & Veenstra, 2013 p. 9). In other words, the original
DP subject contains the cleft clause that Mary saw, which then moves to the end of the
sentence. In the movement, the cleft clause has been taken outside the NP that
contained it (extraposed); the original DP subject is then spelled out as it. Note that the
S-structure is the final sentence.
In the extraposition theory, because the cleft clause is connected with the cleft
pronoun, not the cleft phrase, it-cleft sentences cannot be reduced to simple, noncopular sentences, such as Mary saw John. The cleft clause that Mary saw must refer to it
and not to John. The extraposition step, in which it is spelled out in the S-structure, is
crucial to this theory.
The expletive approach: In linguistics, an expletive pronoun is a non-referential
element such as it or there, which doesn’t refer to anything. The expletive it is different
from a referential it, as in (12), where the pronoun it refers specifically to a noun
elsewhere in the sentence (in this case, car). In the expletive theory of clefts, it is not
important in interpreting the sentence. In a sentence such as (13) (Patten, 2012 p. 6), for
example, Jespersen (1937, as cited in Patten, 2012) claims that the purpose of it is is to
put focus on the cleft phrase; the cleft pronoun is not even present in the base structure.
This differs from his earlier (1927) proposal of the extraposition analysis. Semantically,
IT-CLEFTS AND THEIR PROCESSING
the it-cleft sentence conveys the same meaning as the non-cleft (14) (Patten, 2012 p. 6),
but with a stronger emphasis.
(12) After the car broke down, Steve had it towed to the repair shop.
(13) It was Frank that complained.
(14) Frank complained.
(Patten, 2012)
(15) Expletive approach (sentence from Hartmann & Veenstra, 2013 p. 17)
8
IT-CLEFTS AND THEIR PROCESSING
9
In (15), unlike in (11), John (the cleft phrase/subject) is one argument of the verb
be and the cleft clause that Mary saw is a separate argument of that verb. The cleft
pronoun it is an expletive that blocks subject raising (Hartmann and Veenstra, 2013 p.
17); the expletive is inserted at S-structure. Since it is an expletive, nominative case
marking of the cleft phrase (John) cannot be assigned as a result of case agreement
between the cleft pronoun and the cleft phrase. Additionally, there is an interpretational
relation between the cleft phrase and the cleft clause, not one that is derived from
movement (p. 18). As in (11), the S-structure is the final sentence structure.
In the expletive approach, the it of the cleft sentence cannot refer to anything else
in the sentence because that cleft pronoun is not present in the initial syntactic structure.
Its purpose is to place emphasis on the cleft phrase. The cleft pronoun’s status as an
IT-CLEFTS AND THEIR PROCESSING
10
expletive is critical to this approach, as is the placement of the cleft clause as an
argument of the copular verb.
Kiss (1998) proposes a variation of the expletive approach; in her analysis, not
only is the cleft pronoun it an expletive, but the copular verb is also an expletive. A
major difference with Kiss’s variation is the existence of a Focus Phrase (FP), which is a
layer of structure she introduces on analogy with languages like Hungarian, in which
focused elements have a specific position they appear in. For the cleft construction, she
says that the cleft constituent (to John in (16)), which is the identificational focus of the
sentence, occupies this position. Another fact that makes Kiss’s analysis different: that is
not a relative pronoun. Instead it is a complementizer which blocks V-movement into F,
so F must be filled by the expletive be (Kiss, 1998 p. 258).
(16) It was to John that I spoke. (Sentence from Kiss, 1998)
IT-CLEFTS AND THEIR PROCESSING
11
In (16), neither it nor be are present in D-structure. The sentence itself is also
different from (11) and (15) in that the cleft phrase is a PP, not a DP. This PP moves, in
the final S-structure, to spec-FP from under the embedded VP (Kiss 1998 p. 259). Again,
note that S-structure is the final sentence.
In Kiss’s variation of the expletive theory, the Focus Phrase is introduced
partially to allow her structure to be applied to sentences in both Hungarian and
English. The cleft constituent fills this position. Her variation also includes that as a
complementizer instead of a relative pronoun, which means that in the initial Dstructure, its base position is higher in the tree and prevents the V-movement into the F
position; therefore be must be inserted as an expletive.
The it-as-subject approach: Reeve (2013) claims that while the extraposition
(specificational) analysis of it-clefts is right in classifying the cleft pronoun it as non-
IT-CLEFTS AND THEIR PROCESSING
12
expletive, and that the cleft clause is a form of a restrictive relative clause, he disagrees
with that approach’s determination that there is an extraposition relation between the
cleft clause and the cleft pronoun (Reeve, 2013 p. 173). He instead presents an analysis
in which the cleft clause’s antecedent is the cleft phrase (what he calls the “clefted XP”),
instead of the cleft pronoun. What he means by stating that the cleft clause’s antecedent
is the cleft phrase is that it adjoins to the cleft phrase (Reeve, 2013 p. 175), as illustrated
in (17).
(17)
(Reeve, 2013 p. 166)
Hartmann and Veenstra (2013) also address the it-as-subject approach. In this
approach, unlike in the expletive approach, an it-cleft sentence cannot be reduced to a
non-cleft/noncopular sentence such as (18) because it (the cleft pronoun) is present in
the sentence from the beginning.
(18)
Mary saw John.
The it-as-subject analysis maintains that the cleft pronoun and the cleft phrase are both
arguments of the copular verb—the cleft pronoun is the subject of the sentence and the
cleft phrase and cleft clause are the predicate (p. 12).
IT-CLEFTS AND THEIR PROCESSING
(19)
13
It is John that Mary saw. (Sentence from Hartmann & Veenstra, 2013 p. 13)
IT-CLEFTS AND THEIR PROCESSING
14
A major difference between the it-as-subject approach and the two previous
analyses is that this theory is distinguished by the linking of the cleft phrase (John in
(19)) with the cleft clause (that Mary saw). This connection does not exist in either the
extraposition or the expletive theory. The arguments of the copular verb are as follows:
in the extraposition approach, the cleft phrase (John) is an argument; in the expletive
theory, both the cleft phrase and the cleft clause (that Mary saw) are arguments; and in
the it-as-subject analysis, the cleft pronoun (it) and the cleft phrase are the arguments of
be.
Trying to determine the syntactic structure of it-cleft sentences is still a topic of
active research. While the analyses vary, however, there are some points of
commonality among them. Firstly, and most importantly, the proposed approaches—
with the exception of Kiss’s, in which that is a complementizer and not a relative
pronoun—all include giving the cleft clause a very similar structure to that of a relative
clause, which involves movement of the relative pronoun from its original position.
Secondly, the expletive, extraposition, and it-as-subject analyses all treat the cleft phrase
as an argument of the copular verb. Thirdly, cleft clauses (20), like relative clauses (21),
are treated as modifiers of the cleft phrase (i.e., milk carton in (20-21)).
(20)
It was a milk carton that I threw away.
[object-extracted cleft clause]
(21)
The milk carton that I threw away was empty.
[relative clause]
These commonalities are useful when studying how people process it-cleft sentences,
because there has been significant research on relative clauses as well as some on cleft
clauses.
IT-CLEFTS AND THEIR PROCESSING
15
The analysis that I prefer is the it-as-subject approach, mainly because it seems the
most straightforward to me. The cleft pronoun is present in the syntactic structure right
from the beginning, and the cleft phrase is an argument of the copular verb.
Additionally, the relative pronoun in the cleft clause has the cleft pronoun (the subject)
as its antecedent. From a teaching standpoint, this analysis is closest to how I explain
the antecedent of relative pronouns when discussing relative clauses with my students.
IT-CLEFTS AND THEIR PROCESSING
16
Chapter III: Processing
a. Relative Clause Processing
There has not been much processing work done on clefts, but because almost all
of the syntactic approaches express the idea that the cleft clause segment of it-cleft
sentences is similar to a relative clause, it is appropriate for this thesis to study the solid
body of research that has been done on how people understand relative clauses. The
most general finding in relative-clause processing has been that subject-extracted
relative clauses, as in (22), are easier for people to process than object-extracted relative
clauses, shown in (23).
(22)
The senator that bothered the reporter caused a big scandal.
(23)
The senator that the reporter bothered caused a big scandal.
(Sentences (1) and (2) from Gordon and Lowder, 2012)
Gordon and Lowder (2012) discuss three theories that are used to explain why
object-extracted relative clauses, in which the head NP is the object of the RC, are more
difficult to process than subject-extracted relative clauses, in which the head NP is the
subject of the RC. These theory groups are Memory/Resource-Based Models,
Semantic/Pragmatic Models, and Frequency-Based Models. Memory/Resource Based
Models explain that the head NP (the extracted element) in a subject relative clauses,
such as senator in (22), is connected to the embedded verb (bothered) that occurs
immediately after that NP. When processing an object relative clause, however, the
reader must keep that head NP in reserve until, after succeeding words, it can be used
to help determine a sentence’s meaning, as in (23), in which the embedded verb occurs
IT-CLEFTS AND THEIR PROCESSING
17
after a second NP (reporter). The more words between the NP1 and the embedded verb,
the more difficult the processing because of the distance the NP has to be kept in
memory.
Semantic/Pragmatic Models present the idea that this difference in processing
between object- and subject- extracted clauses can actually be reduced when a semantic
relationship exists between the important NPs and the verb’s (rescued) action, as in (24),
as opposed to NP1 and NP2 being only arbitrarily related to the verb’s (detested) action,
as in (25). The fireman rescued the robber in (24), so both the robber and the fireman
have a relationship with the embedded verb.
(24)
The robber that the fireman rescued stole the jewelry.
(25)
The robber that the fireman detested stole the jewelry.
(Gordon and Lowder, 2012)
Frequency-Based Models maintain that people understand sentences better if those
sentences are written in more routinely encountered structures. Studies have proven
that subject-extracted relative clauses occur more often in English than object-extracted
ones do, so their meanings are more easily determined (Gordon and Lowder, 2012). It
is possible that all three of these theories identify relevant factors in the processing of
sentences containing object-extracted relative clauses.
Warren and Gibson (2005) turn to the processing of cleft sentences; they
concentrate on inter-word dependencies and how that affects how easily people process
complex sentences. In the introduction of their paper, they review the structure of
IT-CLEFTS AND THEIR PROCESSING
18
relative clause sentences like (26) and (27), in which the relative clauses (that the senator
attacked and that attacked the senator) both modify the first NP, reporter.
(26)
The reporter that the senator attacked disliked the editor.
(27)
The reporter that attacked the senator disliked the editor.
(Warren and Gibson, 2005)
The only difference between these two examples is the word order in the relative clause.
As in Gordon and Lowder’s Memory/Resource-Based Models theory, Warren and
Gibson discuss how, when people read sentences, they must remember words that
occur early in the sentence as they encounter new words further on. When, as in (27),
the integration is between consecutive words, rather than across the NP the senator, as in
(26), sentence processing occurs more quickly.
Warren and Gibson go on to question if there is a difference in the processing
time for object-extracted it-cleft sentences when there’s a difference in the type of NPs
the sentence contains (they do not test any sentences in which NP1, whether pronoun,
name, or description, is nominative case). In the experiment conducted, the sentences
contain the more difficult object-extracted clefts only: the cleft NP—NP1—is accusative
case; the subject of the cleft clause—NP2—is nominative case; see (28).
(28)
It was (the lawyer/Patricia/you) who (the businessman/Dan/we)
avoided at the party.
(Sentence (4) from Warren and Gibson, 2005 p. 757)
They determined that varying the NP type among pronouns (e.g., you), first names (e.g.,
Bill), and definite descriptions (e.g., the baker) in both the cleft phrase and the cleft clause
IT-CLEFTS AND THEIR PROCESSING
19
affected the reading time of sentences. They also discovered that the sentences which
had pronouns in the NP2 position were read faster than when the NP2 was a name,
which in turn were read faster than when the NP2 was a description. Additionally,
sentences with pronoun-pronoun conditions were read faster than sentences with
name-name or description-description conditions. These faster reading times are
because it is easier during structure building to integrate across pronouns than names
and across names than descriptions (Warren and Gibson, 2005 p. 762). This is due to
what they call referential processing; their theory is that pronouns are the most
accessible referents when people are determining a sentence’s meaning, first or famous
names come next, and descriptions are the least accessible (Warren and Gibson, 2005 p.
754).
Warren and Gibson also mention the similarity-based integration hypothesis, on
which processing is harder if NP1 and NP2 are of the same type (e.g., It was you who we
avoided at the party, in which NP1 and NP2 are both personal pronouns) than if they are
of different types (e.g., It was the lawyer who Dan avoided at the party, in which NP1 is a
definite description and NP2 is a first name). This theory also helps to explain why
sentences with object-extracted clauses can be more difficult to process than subjectextracted ones. As already discussed, in the object-extracted structure, people must hold
NP1 in memory across NP2, and if NP2 is of the same type, NP1 doesn’t remain clear in
memory.
The research done by Gordon and Lowder and Warren and Gibson helped me to
anticipate how participants in my experiment would rate it-cleft sentences with object-
IT-CLEFTS AND THEIR PROCESSING
20
extracted clefts as compared to the same type of sentences that had subject-extracted
clefts. Their theories also influenced the construction of my experiment sentences to
contain only personal pronouns as the cleft phrase, and, in the case of sentences using
object-extracted clauses, to use a definite description as the NP in the cleft clause.
Additionally, the omission of the who vs. whom factor in Warren and Gibson’s
experiment sentences (they used only who, even though all of their sentences contained
object-extracted cleft clauses) made me more curious about how participants would rate
sentences that followed the prescriptive rules regarding the use of the relative pronoun
whom.
b. Pronouns and Prescriptivism
There are two sentence positions that can contain pronouns after the
initial it in cleft sentences: the cleft phrase may be a personal pronoun instead of a noun,
and the relative pronoun in the cleft clause will be who, whom, that, or which. Prescriptive
rules tell us that the cleft phrase pronoun is supposed to be a nominative pronoun, as in
(29), though an accusative pronoun, as in (30), is widely accepted. Prescriptive rules also
tell us that if the cleft clause is a subject-extracted one, the nominative who should be
used (29) and that the accusative whom is incorrect in subject-extracted clauses (31).
Other issues of prescriptivism for the cleft clause pronoun involve the acceptability of
who in an object-extracted clause and the caseless that used with animates when it
should, according to prescriptive rules, only be used with inanimates.
(29)
It is he who writes bestselling novels.
(30)
It is him who writes bestselling novels.
IT-CLEFTS AND THEIR PROCESSING
(31)
21
*It is he/him whom writes bestselling novels.
I will summarize some relevant work on the processing and use of pronouns,
especially those that have prescriptive issues. MacKay (1980), for example, looks at the
issue of pronouns from a psychological perspective. His study investigates the use of
the pronoun he to mean “he or she.” The original prescriptive use of he in this manner
began over 250 years ago; this use continues and the attempts to analyze and defend
this prescription include:
(a) the pronomial surrogate assumption, in which pronouns stand for
antecedents and contribute no new meaning; he indicates “person,” but
doesn’t exclude women;
(b) the semantic-flexibility assumption, which states that a word’s meaning is
highly flexible and includes “special-purpose” definitions. People have no
difficulty in understanding/learning these definitions; therefore, Burgess
(MacKay p. 445) declared that his use of he was neutral and those who
thought otherwise were forcing “chauvinistic sex onto the word”; and
(c) the context assumption, where the prescriptive he resembles an ambiguous
noun that has several semantic duties but can easily be interpreted in context.
So people don’t confuse he in contexts referring to “people” with contexts
where he specifically refers to “a man.” (MacKay p. 445).
MacKay had subjects read paragraphs that used he to refer to neutral antecedents
(e.g., person) and answer multiple-choice questions assessing comprehension of
prescriptive he and antecedents. This experiment also examined how participants
IT-CLEFTS AND THEIR PROCESSING
22
understood a novel pronoun (e, E, or tey) read for the first time without explanation
under the same conditions as subjects reading the prescriptive he. The results
contradicted all three assumptions. For the pronomial surrogate assumption, 80% of
the subjects in 75% of the trials understood neutral antecedents of he as male rather than
as male or female. In the case of the semantic-flexibility assumption, it was discovered
that the maleness of the prescriptive he is so ingrained in semantic memory that it can’t
be displaced by special-purpose meaning. The context assumption was contradicted
because the results showed that context is not helpful in resolving the prescriptive he’s
ambiguity; listeners wait to hear he or she when a sex-indefinite noun (such as child) is
used, since there is no generic pronoun for third-person singular. This illustrates a
situation in which even though prescriptive grammar mandates a specific use of a
particular word (in this case, the neutral understanding of the prescriptive he pronoun),
people do not process the word in that prescribed manner.
Another way in which prescriptive grammar can cause pronoun problems
involves what Wolfram and Schilling-Estes (1998) call style-shifting. An example of this
shifting is changes in the formality of speech: less formal to more formal (dropping “g”
from “ing,” gonna vs. going to, double negatives) or vice-versa. In the shift from less
formal (e.g., conversational grammar, for example) to more formal language (e.g.,
writing for an English class), people can exhibit hypercorrection—they can make
mistakes based on a misunderstanding of the rules of formal grammar. One example of
this hypercorrection is the incorrect use of pronouns. For example, people often use a
subject pronoun instead of the grammatically correct object pronoun in coordinations:
IT-CLEFTS AND THEIR PROCESSING
23
they will say or write She’s going with you and I instead of She’s going with you and me. Or
they will use a reflexive pronoun instead of an object pronoun: Bring any project ideas to
myself instead of Bring any project ideas to me. This phenomenon can be extended to the
use of who versus whom. People may be aware that there is a difference between the two
pronouns, but they may misuse them because they don’t understand the formal pattern
that applies to their use.
The misuse of pronouns is not a new phenomenon, nor does it seem to be
restricted to any particular group of people. Robert J. Geist, in “Professors’ English”
(1952), examined the use of ungrammatical words and phrases in everyday language.
Geist did not believe that educated people (especially his Ph.D. colleagues) would use
awkward and/or ungrammatical phrases, but upon listening more closely to his peers,
he noticed such errors as without you and I being on campus and for my brother and I (p.
16)—both instances in which the correct pronoun should have been me, not I. He also
gave examples of errors occurring in textbooks and the nineteen solecisms covered in
Knickerbocker (1950). These included mistakes in using the future tense will when only
the present tense is necessary, has got, go slow (missing the adverbial ly), and It is me and
Who did you meet. Geist pointed out that even in written sentences in such well known
works as Hawthorne's The Scarlet Letter, the Encyclopedia Britannica, and Webster's New
Collegiate [Dictionary], prescriptive rules have not always been followed. Such examples
from almost a century ago bring home the fact that many prescriptive rules have been
inconsistently adhered to for a long time, not only by those without formal education,
but also among educated groups.
IT-CLEFTS AND THEIR PROCESSING
24
In addition to the gender pronoun issue raised by MacKay’s research, I
discovered articles from each decade of the past 70 years that raise the case issue for
relative pronouns: the expectation of the demise of the accusative whom in favor of
using who for both nominative and accusative was discussed in all seven articles. J. T.
McM., of the N.C.T.E. Committee on Current English Usage, wrote in 1945 that even the
then-current Webster’s New International agreed that using who instead of whom as an
object of either a preposition or a verb was common and “still found in good writers”
(McM. p. 104). Miller (1957) had several different ideas involving grammar rules,
including the who/whom question: English teachers weren’t enforcing correct grammar;
there were many people who not only weren’t familiar with prescriptive grammar
rules, but also had no desire to learn them; and there were groups of people in New
York City, among them, editors and radio announcers, who should know the proper
uses of who and whom but wouldn’t always use them correctly (Miller p. 136). He came
to the conclusion that he might teach the “proper” use of whom, but that dropping whom
in everyday conversation would not be a bad thing. Frank (1962) referenced Noah
Webster in 1783 declaring that whom was a useless pronoun, especially when people
attempted to use it at the beginning of a sentence. Who did you vote for?, for example,
was preferred over Whom did you vote for?, mainly because the belief was that people
couldn’t tell why whom should be used instead of who in that sentence structure. But if
Ernest Hemingway, in 1940, had agreed with Webster, English majors and scholars
would be studying For Who the Bell Tolls, which doesn’t have quite the same ring to it as
the actual, grammatically correct title.
IT-CLEFTS AND THEIR PROCESSING
25
The 1970s and 1980s were no different; Long (1975) also complained that while
the distinct nominative and objective forms of the personal pronouns (e.g., I vs. me) and
when to use them are relatively well understood, the problem of when to use whom
persisted. He went so far as to suggest that who and whoever should be the “commoncase” forms, just as that is a caseless relative pronoun, and recommended dropping
whom and whomever (Long p. 198). In accordance with earlier writers who believed that
whom’s disappearance is imminent, Redfern (1981) reveals that he also would like to
witness “The Death of Whom” (the title of his article); however, unlike earlier writers,
he does admit that it will probably still be present in the language for at least another
100 years (p. 83). Soles (2005) is even more emphatic in his desire to see whom disappear;
he would like to see it banished from our vocabulary and its spot taken over by who,
just as thee and thou were replaced by you. All of these scholars and educators believe
that whom will disappear, though they vary on the length of time it will take before that
actually occurs, and if it should or will happen in both conversational and formal
English.
Current textbooks also confirm that even though prescriptive rules are still
taught, there is more acceptance in formal writing of certain pronoun errors. For
example, both the tenth (2007) and eleventh (2014) editions of The Blue Book of Grammar
and Punctuation (Jane Straus, et al.), include the specific rule that the relative pronoun
who should be used in referring to people and that and which should be used when the
antecedents are groups or things. The tenth edition declares that who (or whom, in the
accusative case) is the only pronoun that should be used in conjunction with people and
IT-CLEFTS AND THEIR PROCESSING
26
only alludes to the fact that the pronoun that may refer to people in answers to practice
quizzes. The eleventh edition, however, directly states in Rule 1 under the “Who, That,
Which” section of the grammar chapter that this “misuse” of pronouns is acceptable (p.
14) even in formal situations. Additionally, an email newsletter sent out to subscribers
in September 2014 from Grammarbook.com (the textbook’s website) gives examples
from authors, editors, and scholars from the 1990s, ‘80s, and ‘60s of the use of that in
place of who, both formally and informally. The newsletter ends with a Bible quote
containing the pronoun that in place of who and declares that while several adjectives
have been used to describe the Bible over the centuries, it’s not likely that “informal”
has been one of them.
Given the “bending” of prescriptive grammar rules, including a greater
acceptance of what were formerly “incorrect” pronouns (It was me instead of It was I, for
example), the predictions of the disappearance of whom, and the tendency of people to
exhibit hypercorrection because they are unsure of which words to use in formal
English (Pete went with John and I to the park instead of John and me), will people still
recognize when correct grammar is presented to them in the cleft structure? Will they
think something is “proper” grammar just because they are unfamiliar with it? Through
the experiment created for this thesis, I hope to find out.
IT-CLEFTS AND THEIR PROCESSING
27
Chapter IV: Experiment
Much of the published syntactic and processing literature within linguistics does
not follow the prescriptive use of who and whom (for example, in (28) above from
Warren and Gibson, the pronoun who should actually be whom because the cleft NP is
accusative case). This caught my attention because I teach these rules. Through this
experiment, which was a sentence rating study, I wanted to attempt to answer the
following questions regarding who and whom: Will college students know how to
recognize the correct use of whom, or will they rate it lower because they don’t know
how that pronoun is formally used? Will they rate versions of the sentences containing
who higher, even if the pronoun beginning the cleft clause should, according to
prescriptive grammar, be whom?
I also was interested in discovering how participants rated the variations
between nominative and accusative personal pronouns (e.g., I/me or they/them in the
cleft phrases). Will the students recognize that, following prescriptive rules, the
pronoun in that position should always be a nominative pronoun, since the cleft
pronoun it is considered a subject pronoun? Or will they rate the sentences with the
accusative pronoun in the cleft phrase position just as high since, according to more
recent grammar rules, the use of It is me or It is them has become more acceptable even
in formal writing (Straus, et al., 2014 p. 8)?
I predict that the sentences which will receive the highest rating are those with
subject-extracted cleft clauses that have an accusative pronoun as the cleft phrase and
IT-CLEFTS AND THEIR PROCESSING
28
use who in the cleft clause (condition subject cleft/accusative/who). I base this prediction on
the following:
•
Even people who are studying the use of language use accusative pronouns
(e.g., me/him) where, prescriptively, nominative pronouns should have been
used (Warren and Gibson’s cleft phrase pronouns, for example [2005 p. 766]).
•
The prescriptively correct uses of who as subject and whom as object have not
been followed (examples exist from as early as 1467), especially in informal
English (Aarts, 1994 p. 71); in fact, Aarts claims that “if whom is still with us, it
is not because it plays a prominent part in the language” (Aarts, 1994 p. 74).
•
A recently-updated grammar textbook acknowledges that even English
scholars have been more accepting of the use of accusative pronouns as the
cleft phrase, narrowing the distinction between formal and casual English
(Straus, et. al., 2014).
I also predict that the sentences with subject-extracted clefts will receive higher
ratings than the sentences with object-extracted clefts. I believe this because subjectextracted clefts are easier for people to process, as discussed in research by Gordon and
Lowder and Warren and Gibson. The Frequency Based Model, for example, has shown
that subject-extracted clauses appear in English much more often than object-extracted
clauses and suggests that this frequency leads to easier understandability.
I also intend to test how people produce clefts. If people are indeed unfamiliar
with the who/whom distinction, I predict the following when participants produce their
own sentences using the it-cleft structure:
IT-CLEFTS AND THEIR PROCESSING
•
29
Participants will misuse whom (i.e., use whom where who should prescriptively
be used) more frequently than using it correctly, or they will avoid the
problem entirely by using that, which has no case assigned, instead.
Further, I predict that there will be patterns in the ratings given by groups of
participants with similar production issues:
•
Those participants who do misuse whom in sentence production will rate the
whom sentences higher than the group as a whole.
•
Those who use that instead of who or whom when producing their own
sentences will rate the experiment sentences higher than the group as a
whole.
Method
Materials. The experiment included 24 it-cleft sentences of the forms seen in (32)
and (33). See Appendix A for the complete set of experiment items. There were 12
sentences with a subject-extracted cleft clause as in (32) and 12 sentences containing an
object-extracted cleft clause as seen in (33).
Each of the 24 items had four conditions, two with nominative personal
pronouns in the cleft phrase position, as in (32a) and (32b), and two with accusative
personal pronouns, as in (32c) and (32d). The cleft clause contained either the
nominative pronoun who as seen in (32a) and (32c) or the accusative pronoun whom as
seen in (32b) and (32d).
(32) a. It was I who completed the assignment.
b. *It was I whom completed the assignment.
IT-CLEFTS AND THEIR PROCESSING
30
c. It was me who completed the assignment.
d. *It was me whom completed the assignment.
(33) a. It was we who the mayor thanked.
b. It was we whom the mayor thanked.
c. It was us who the mayor thanked.
d. It was us whom the mayor thanked.
A post-questionnaire section, which was the same for all 12 versions, provided the
participants with non-cleft sentences, as in (34a) through (37a), and an it-cleft fill-in-theblank new sentence beginning that they need to complete for each, as in (34b) through
(37b).
(34)
a. Matilda ran for president of the student body.
b. It was Matilda ____________________
(35)
a. The car needed a new engine.
b. It was the car _____________________
(36)
a. Justin dated Sara last year.
b. It was Sara _______________________
(37)
a. The store sold oranges this winter.
b. It was oranges ____________________
The sentences in (34) and (35) were geared towards the creation of subject-extracted
cleft clauses; (36) and (37) were designed to produce object-extracted completions.
Subjects. One-hundred eighteen students from South Georgia State College
(SGSC) and seven students at Morehead State University (MSU) completed the
IT-CLEFTS AND THEIR PROCESSING
31
questionnaire. The SGSC students were undergraduates in either remedial English or
English Composition classes and were given extra credit points by their instructors for
their participation. The MSU students were recruited from psychology classes and were
paid $10 for their participation.
Procedure. The rating questionnaire contained the 24 experiment sentences along
with 24 and 20 sentences respectively from two unrelated experiments, 10 it-cleft filler
sentences, and 26 non-cleft fillers, all varying in grammaticality, for a total of 104
sentences. The filler it-cleft sentences differed from the experiment sentences in that the
cleft phrase contained a noun instead of a personal pronoun and some of them were
outright ungrammatical. There were twelve versions of the questionnaire, each
containing 104 sentences to be rated. The items were counterbalanced so that each
participant saw only one version of each experimental item and saw an equal number
of items in each condition over the experiment. The items appeared in one of twelve
pseudo-randomized lists such that no consecutive items were of the same type.
Participants were given a paper questionnaire and asked to rate the sentences on
a scale from 1 (ungrammatical or unnatural) to 7 (natural and understandable). The
majority of the students were in a classroom environment; the experiment was not
timed. The instructions were read aloud to participants, as well as being included as the
first page of the questionnaire packet. Most students completed the questionnaire
within 20 minutes; the ones who took longer were finished within 30 minutes. As the
completed questionnaires were returned, they were each marked with a different
subject number.
IT-CLEFTS AND THEIR PROCESSING
32
Results
Sentence Ratings
Table 1: Average ratings of all participants
nominative
accusative
who
whom
who
whom
object cleft
4.04
4.16
4.36
4.45
subject cleft
4.47
4.23
4.54
4.15
The rating results were analyzed using analyses of variance (ANOVAs) by
subjects and by items. For the subjects analysis, all three factors (syntactic type of cleft,
pronoun case, and relative pronoun case) were within subjects. For the items analysis,
pronoun case and relative pronoun case were within items, and syntactic type of cleft
was between items.
The main effect of syntax (i.e., are the ratings for all subject clefts different from
the ratings for all object clefts) was not at all significant by items (p = .53) and marginal
by subjects (F1(1,124) = 3.45, p = .065).
The main effect of pronoun case (i.e., are the ratings for all clefts with nominative
personal pronouns different from the ratings for all clefts with accusative pronouns?)
was significant by subjects (F1(1,124) = 5.36, p = .022), and marginal by items (F2(1,22) =
2.98, p = .098). Overall, clefts with accusative personal pronouns were rated slightly
higher than those with nominative pronouns. Additionally, the type of it-cleft sentence
that received the highest average rating was the subject cleft/accusative/who, as in (38); the
object cleft/accusative/whom, as in (39), was rated the second-highest among those with
accusative pronouns.
IT-CLEFTS AND THEIR PROCESSING
(38)
It was me who gave the speech.
(39)
It is him whom the lion injured.
33
The main effect of who/whom (i.e., are all of the ratings for clefts with who
different from all the ratings for clefts with whom) was not significant by items or
subjects, p’s > .10.
The interaction between syntax and pronoun case (i.e., are object clefts with
nominative personal pronouns different from subject clefts with nominative pronouns,
and are object clefts with accusative personal pronouns different from subject clefts
with accusative pronouns?) was significant by subjects (F1(1,124) = 6.85, p = .01) and
marginal by items (F2(1,22) = 3.56, p = .073). These results show that subject clefts get
higher ratings with nominative pronouns, as in (40), and object clefts get higher ratings
with accusative pronouns, as in (41). This could be evidence of a matching effect—i.e.,
when subject cleft clauses (in which nominative pronouns are required by prescriptive
rules) are paired with nominative cleft phrase pronouns, the sentence gets a higher
rating than when an accusative cleft phrase pronoun is used.
(40)
It was she who painted the portrait.
(41)
It was her whom the teacher sent to the principal’s office.
The interaction between syntax and relative pronoun case (i.e., are ratings for
subject clefts with whom vs. who different from ratings for object clefts with whom vs.
who) was significant by subjects (F1(1,124) = 14.01, p < .001) and significant by items
(F2(1,22) = 7.28, p = .013). These results show that subject clefts were always better with
who than with whom, as in (42) and (43); object clefts were always better with whom than
IT-CLEFTS AND THEIR PROCESSING
34
with who, as in (44) and (45). This appears to be another matching effect—i.e., when the
relative pronoun in the cleft clause is nominative case and the clause is a subjectextracted one, the sentence received a higher rating than one in which the relative
pronoun was accusative case.
(42)
It was I/me who gave the speech.
(43)
*It was I/me whom gave the speech.
(44)
It is he/him whom the lion injured.
(45)
It is he/him who the lion injured.
The interaction between pronoun case and relative pronoun case (i.e., are ratings
for clefts with nominative pronouns and who different from those for clefts with
accusative pronouns and who, and the same for nominative/whom and
accusative/whom) was not significant by subject or items, p’s > .4. The interaction
between syntax, pronoun case, and relative pronoun case (i.e., are the combinations of
all three factors different from each other) was not significant by subjects or items, p’s >
.4.
Let us examine some specific conditions and how the significant effects show up
in their ratings. The first set of conditions discussed are the object-extracted clefts; the
second set is the subject-extracted cleft conditions.
Object-extracted clefts:
Condition 1 consists of sentences of the object cleft/nominative/who structure, for
example, It was we who the mayor thanked. This type of sentence (using who where whom
is grammatically correct because it’s an object-extracted cleft) is becoming more
IT-CLEFTS AND THEIR PROCESSING
35
commonly used (even by scholars—see example (28) on page 17 taken from Warren and
Gibson; their experiment sentences used all object-extracted clefts but used who as the
relative pronoun). This condition received an average rating of 4.04, the lowest of all,
even though it is acceptable in terms of changing grammar rules. The rating is low
because this type of sentence is an object cleft with two nominative pronouns, so there
is no matching of the syntax with the case of either pronoun. Also, object clefts are
known to be harder to process.
Condition 2 and Condition 3: Condition 2 involves object-extracted cleft sentences
with nominative cleft phrase pronouns but an accusative relative pronoun (It was we
whom the mayor thanked.); Condition 3 has the reverse: an accusative cleft phrase
pronoun and a nominative relative pronoun (It was us who the mayor thanked.) These two
conditions received intermediate ratings (4.16 and 4.36 respectively) because one of the
pronoun types (but not both) matches the cleft syntax.
Condition 4 is made up of sentences of the object cleft/accusative/whom structure,
such as It was us whom the mayor thanked. This is one of the object-extracted sentence
types that, according to prescriptive grammar, is correct in using whom to begin the cleft
clause. Using an accusative cleft phrase pronoun has become common and acceptable.
Rating: 4.45, the highest of the object-extracted cleft conditions; both pronoun types
match the cleft syntax.
Subject-extracted clefts:
Condition 1 and Condition 3 both contain sentences that have subject-extracted
clefts and nominative relative pronouns (e.g., It is they/them who bought the bikes.). These
IT-CLEFTS AND THEIR PROCESSING
36
two sentence types were the best in terms of prescriptive grammar rules, and received
the two highest ratings (4.47 and 4.54, respectively).
Condition 2 consists of subject-extracted cleft sentences with a nominative cleft
phrase pronoun and accusative relative pronoun (*It is they whom bought the bikes.)
Because of the presence of whom at the beginning of the cleft clause, this type of
sentence is ungrammatical. Yet it did not receive the lowest average rating; in fact, the
rating (4.23) is higher than the object cleft/nominative/who and object cleft/nominative/whom
sentences, both of which are grammatically acceptable. It is possible that this higher
rating could be due to hypercorrection.
Condition 4 involves subject-cleft sentences that have both an accusative cleft
phrase pronoun and relative pronoun (*It is them whom bought the bikes.) This is the other
sentence type that, like the subject cleft/nominative/whom condition, is ungrammatical
because of the use of whom to begin the cleft clause. Participants, as shown by the low
average, recognized the ungrammaticality, though it is surprising that the average was
not even lower, at the 1 or 2 level. The average rating for these sentences was 4.15, the
lowest-rated subject-extracted clause condition. The lack of a matching effect (i.e., a
nominative cleft phrase pronoun with an accusative relative pronoun) could also
explain the low rating.
Fill-in Cleft Sentences
The responses of the fill-in portion of the questionnaire were recorded to include
the verb of the cleft clause. One hundred and twenty-five (125) participants responded.
Sentences (46) through (49) are the grammatical responses to the prompts.
IT-CLEFTS AND THEIR PROCESSING
37
(46)
It was Matilda who ran for president of the student body.
(47)
It was the car that/which needed a new engine.
(48)
It was Sara whom/that Julian dated last year. (This would keep the objectextracted structure, as opposed to switching to subject-extracted structure
as in It was Sara who/that dated Julian last year.)
(49)
It was oranges that/which the store sold this winter. (This would also
keep the object-extracted structure, as opposed to It was oranges that were
sold by the store this winter.)
Table 2: Actual sentence responses through critical verb
Sentence Prompt
It was Matilda …
It was the car …
`It was Sara …
Completions
who ran
that ran
*whom ran
*whom was voted
of the student body, who ran
who had run
who is running
?who the student body let run
Other1
that needed
which needed
that needs
*who needed
*who needs
engine that needed to be renewed
that got
that had a blown engine
that required
that the new engine was needed for
that was in need of
Other1
who dated
that dated
*whom dated
whom Julian dated
*who Julian dated
Number of Responses
80
20
15
1
1
1
1
1
5
102
4
2
1
1
1
1
1
1
1
1
9
44
25
13
13
7
IT-CLEFTS AND THEIR PROCESSING
38
that Julian dated
6
and Julian who dated
2
and Julian that dated
1
that had dated
1
that went out with
1
that went with
1
who broke up with
1
Other1
10
It was oranges …
that the store sold
45
that were sold
38
that were being sold
5
?that sold
3
which were sold
2
that the store was selling
1
which the store sold
1
which were being sold
1
Other1
29
1”Other” includes responses with misspellings, non-cleft structure, and/or grammatical
errors.
For the Matilda and car sentences, in which the intention was the production of
subject clefts, the majority of the participants gave the grammatical responses; there
were only 19 sentence completions in which the pronouns who and whom were
incorrectly used. There was only one instance where the sentence structure for It was
Matilda … was completed with an object-extracted clause instead of the expected
subject-extracted clause. In the Sara and oranges sentence completions, participants
chose more frequently to shift from the intended object-extracted clefts to subjectextracted clefts. For the Sara sentences, 71% of participants provided subject rather than
object clefts, and for the oranges sentences, 39% of them not only produced subject clefts.
The It was Sara … completions are not necessarily surprising, since the idea of two
people dating is conveyed equally well whether Julian dated Sara or Sara dated Julian.
Creating a subject-extracted cleft for the It was oranges … completion, however, involved
IT-CLEFTS AND THEIR PROCESSING
39
more thought for the participants to make that change: the verb, in most cases, had to be
changed from active to passive (since oranges can’t sell something, but the store can) in
order to create an understandable sentence.
This is not only consistent with the research (Gordon and Lowder, 2012) that
people process subject-extracted clefts more easily and quickly than object-extracted
ones, but it also extends the idea further to demonstrate that the participants made a
great effort to use subject clefts over object clefts, even though the prompts for the Sara
sentence and the oranges sentence should have led the subjects to use object-extracted
clefts for the sentence completions. This change in the oranges sentences, while less in
terms of percentage, is more important because of the need for the verb change in order
to make the sentence comprehensible.
We can also look at the fill-ins for the overall use of different relative pronouns,
as in Table 3.
Table 3: Breakdown of cleft clause relative pronoun responses
Original
Sentence
Matilda ran for
student body
president.
The car needed
a new engine.
Julian dated
Sara last year.
The store sold
oranges this
winter.
Provided
Prompt
It was
Matilda …
Who
Whom
That
Which
Other
84
(67%)
16
(13%)
20
(16%)
0
(0%)
5
(4%)
It was the
car …
It was Sara
…
It was
oranges …
2
(2%)
7/47*
(6%/38%*)
0
(0%)
0
(0%)
13/13*
(10%/10%*)
0
(0%)
110
(88%)
6/29*
(5%/23%*)
46/46*
(37%/37%*)
4
(3%)
0
(0%)
1/3*
(1%/2%)
9
(7%)
10
(8%)
29
(23%)
*Participants changed sentence structure from expected object-extracted cleft to subjectextracted cleft.
IT-CLEFTS AND THEIR PROCESSING
40
Participants predominantly adhered to the prescriptive use of who in the It was
Matilda … completion (67%). A little over one-third of the participants (38%) also used
who in a grammatically correct manner for the It was Sara … completion, though that
use was correct only because those participants wrote a subject-extracted clause
completion for the Sara prompt instead of the looked-for object-extracted clause
completion, which would prescriptively have required whom. So, overall, the
prescriptive use of the pronoun who to refer to people was mostly followed, though a
significant percentage of participants (16% and 28%, respectively) used the caseless that
to refer to Matilda and Sara. Additionally, the use of whom, even in sentences where it
would have been expected, such as It was Sara whom Julian dated last year, was rare, with
only 20% of the participants using it in the Sara sentence—and half of those participants
used it incorrectly (*It was Sara whom dated Julian …).
For the sentences involving things as the cleft phrase (It was the car… and It was
oranges …), the prescriptive use of that for inanimate objects was overwhelmingly
followed (88% and 74%, respectively); only a small percentage of participants used
which (3% for each sentence completion). The misuse of who when referring to a thing
occurred minimally (2% of participants) in the It was the car … completion, and not at all
in the sentence containing oranges. Whom was not used in either the car or the oranges
sentence completions. The rule that specifies the use of that for inanimates presented no
problems for the participants, as opposed to the variety of who/whom/that responses for
the sentences involving people (animates).
IT-CLEFTS AND THEIR PROCESSING
41
Rating Patterns of Sub-groups of Participants
Since we have fill-in data for the same participants who completed the rating
task, we can examine subgroups of participants that performed similarly in the fill-in
task.
Table 4: Average ratings of participants misusing whom in fill-in clefts
nominative
accusative
who
whom
who
whom
object cleft
4.24
4.42
4.25
4.29
subject cleft
4.32
4.33
4.13
4.20
There were 23 participants who misused the pronoun whom in the fill-in cleft sentence
completions. When looking at the average ratings of those participants for similarly
constructed sentences (subject cleft/nominative/whom), such as (50), the ratings were
higher for this subgroup than the average rating of the entire group of participants (4.33
as compared to 4.23).
(50)
*It was she whom painted the portrait.
Table 5: Average ratings of participants NOT misusing whom in fill-in
clefts
nominative
accusative
who
whom
who
whom
object cleft
4.01
4.10
4.39
4.49
subject cleft
4.51
4.21
4.63
4.14
When comparing the subgroup of those who used whom correctly in the fill-in
clefts with those who misused whom, as seen above, the difference in the averages is
notable, especially in the case of sentences such as (51) (4.49 vs. 4.29), which follows the
IT-CLEFTS AND THEIR PROCESSING
42
prescriptive grammar rule that whom should be used to begin an object-extracted cleft.
Also, the grammatically correct sentence shown in (52) was rated very high by the
participants who used whom correctly (4.63) and very low by the participants who
misused whom (4.13). Similar results (4.51 vs. 4.32) are seen for the grammatically
“perfect” sentence type shown in (53). The participants who used whom correctly also
rated both ungrammatical sentence types, as in (54) and (55), appreciably lower than the
participants who misused whom (4.21 vs. 4.33 and 4.14 vs. 4.20).
(51)
It was her whom the dog followed in the park.
(52)
It was me who gave the speech.
(53)
It was I who gave the speech.
(54)
*It was I whom gave the speech.
(55)
*It was me whom gave the speech.
Table 6: Average ratings of participants using only that in fill-in clefts
nominative
accusative
who
whom
who
whom
object cleft
4.46
4.35
4.77
4.91
subject cleft
4.45
4.39
4.50
4.29
There were 14 participants who used that in all four sentence completions, as in
(56). These participants avoided having to make a case decision between who
(nominative case) and whom (accusative case) when writing their sentence completions,
but their ratings for the experiment sentences, all of which used who and whom, are
appreciably higher than the overall average ratings for all conditions except for
sentences such as (57) (4.47 vs. 4.45) and (58) (4.50 vs. 4.54).
IT-CLEFTS AND THEIR PROCESSING
43
(56)
It was Matilda that ran for president of the student body.
(57)
It is they who play soccer on the weekends.
(58)
It is them who play soccer on the weekends.
Table 7: Average rating of participants changing object to subject cleft
nominative
accusative
who
whom
who
whom
object cleft
3.88
4.08
4.03
4.32
subject cleft
4.24
4.05
4.24
3.98
There were 100 participants who created subject-extracted clefts, as in Table 2, for
either or both of the two sentences that were geared towards the creation of objectextracted cleft completions. The average ratings of the participants who made the
object-to-subject cleft change in the oranges sentence were lower for all conditions than
the average ratings of all participants. This was surprising; I expected that this
subgroup’s ratings for all conditions of the subject-extracted cleft sentences would be
higher than the overall averages since these participants seemed to favor the subjectextracted cleft structure. The lower average ratings for the object-extracted cleft
sentences, however, were not unexpected as the participants, again, avoided the
construction of object-extracted clefts in their sentence completions.
Several other minor patterns, which were interesting but not statistically
significant, can be found Appendix B.
IT-CLEFTS AND THEIR PROCESSING
Chapter V:
44
Discussion
The first prediction, which stated that sentences with the accusative pronoun
cleft phrases (e.g., me, them, her) and using who in the subject-extracted cleft clause (e.g.
It was me who gave the speech) would receive the highest ratings, was confirmed. These
results show that the use of an accusative cleft phrase pronoun did sound more “natural
and understandable” to the participants than using a nominative cleft phrase pronoun;
they also demonstrate that college students did indeed recognize that who (a subject
relative pronoun), should—according to prescriptive rules—be used instead of whom
(an object pronoun) to introduce the subject-extracted cleft clause. This condition also
has, specifically, matching of the case of the nominative relative pronoun (who) with the
syntactic position of the cleft.
The second prediction, which stated that the sentences containing subjectextracted clefts would receive higher ratings than those sentences using object-extracted
clefts, was confirmed in three of the four conditions: subject cleft/nominative/who, subject
cleft/nominative/whom, and subject cleft/accusative/who. The sentences of the fourth
condition, subject cleft/accusative/whom, received the lowest rating of the subjectextracted cleft types, and the average rating was appreciably lower than the
corresponding object cleft/accusative/whom sentence because of the mismatch of the
syntax and cases. The other ratings are generally consistent with previous research that
shows object clefts are more difficult to process than subject clefts. Still, the effect was
weak, perhaps due to the other conditions making some subject clefts not fully
grammatical.
IT-CLEFTS AND THEIR PROCESSING
45
Subject-cleft sentences with nominative cleft phrase pronouns (I/we) and a
nominative relative pronoun (who) received high ratings, as did object cleft sentences
that had accusative cleft phrase pronouns (me/us) and an accusative relative pronoun
(whom). The nominative/nominative and accusative/accusative matching effects are
prevalent; these sentences apparently sounded more “natural and understandable” to
participants and also reinforce the finding that participants clearly recognize that the
accusative whom is not appropriate when used in a subject-extracted cleft sentence.
In the fill-in sentence completions, the prediction that participants would misuse
whom or use that in their sentence completions was not confirmed in the case of the It
was Matilda … completion. Only 13% of participants misused whom in that sentence
completion, and 20% used that, but 67% of the participants used who to complete the itcleft structure. This suggests that the majority were secure in their knowledge of the
prescriptive use of who, and also that they did not, for the most part, feel the need to use
the caseless that.
The results for the misuse of whom in the It was Sara … completion need to be
examined because 48% of the participants provided a subject-extracted cleft clause
containing either who or whom instead of the expected object-extracted clause. Of that
48%, 10% misused whom in that cleft clause, thereby creating an ungrammatical
sentence (*It was Sara whom dated Julian). Overall, 38% of the participants either misused
whom or used that as compared to 44% who chose who, which is grammatically correct
in the case of a subject-extracted clause and at least accepted, if not preferred, in
introducing an object-extracted clause.
IT-CLEFTS AND THEIR PROCESSING
46
The prediction that those participants who misused whom would give the
sentences containing whom high ratings was generally confirmed, since it was true in
three of the four whom conditions: object cleft/nominative/whom, subject
cleft/nominative/whom, and subject cleft/accusative/whom. For the fourth condition (object
cleft/accusative/whom), however, this same group of participants rated those sentences
notably lower than those participants who used whom correctly, which could be because
they have a general bias towards the nominative case (for example, they might
erroneously “fix” a correct phrase like for John and me to for John and I).
The results from the subgroup whose members used that for all of their fill-in
completions confirm the prediction that this group would assign higher ratings to the
experiment sentences: in 6 of the 8 conditions, those ratings were higher than the ratings
of all participants, and in the other 2 conditions, the ratings were only slightly lower
than the average ratings of all participants (4.50 vs. 4.54 and 4.45 vs. 4.47). Their
avoidance of using who or whom in the sentence completions and the high ratings they
assigned to even ungrammatical sentences (especially both subject cleft/whom
conditions) suggests that they chose that for the sentence completions because, like the
group that misused whom, they do not have a grasp of the rules for using that pronoun.
The results show that people gravitate to the use of accusative personal pronouns
(e.g., me, them) instead of the prescriptively required nominative personal pronouns
(e.g., I, they) in it-cleft sentences; this use has also become more accepted by grammar
experts in recent years. The experiment also confirmed earlier research that subjectextracted clefts are more easily processed than object-extracted clefts. Whether in
IT-CLEFTS AND THEIR PROCESSING
47
subject- or object-extracted cleft sentences, the matching effects of
nominative/nominative and accusative/accusative were clearly emphasized by the
rating results. The misuse of whom in the fill-in sentences was not as prevalent as
expected, though the average rating for sentences containing whom was much higher
with those who did misuse whom than the rating of those who used whom correctly in
the fill-in portion of the experiment. For those participants who used that in the fill-ins,
the confirmation of the expectation that this group would rate all of the experiments
higher than the whole group of participants shows that there are still a good number of
students who don’t understand the rules governing the use of who and whom. The
information gained from these results gives me some clear ideas of how I might teach
English grammar and composition classes in the future.
Going Forward
The main thing I learned through this project is what versions of cleft sentences
students prefer (e.g., why object clefts are more difficult to process and why students
might exhibit hypercorrection when dealing with who versus whom). In light of this
better understanding, there are several results that showed me areas in which I may
need to be flexible about prescriptive grammar rules. There are, however, others in
which I feel justified in enforcing those rules.
For example, the main effect of pronoun case showed that the it-cleft sentences
containing accusative pronouns (me/us) rated higher than those with nominative
pronouns (I/we). Additionally, many researchers (such as Warren and Gibson, 2005) use
accusative cleft phrase pronouns in their own experiment sentences without comment,
IT-CLEFTS AND THEIR PROCESSING
48
and grammar textbooks teach more acceptance of the accusative cleft phrase pronoun.
Perhaps adherence to the It is I structure does not need to be as strongly emphasized in
formal writing. I would, however, want to make sure that the acceptance of an
accusative pronoun where a subject pronoun should go in the cleft structure did not
carry over into the increased use in non-cleft sentences such as Me and her are going on a
date tonight, which I have heard often in my classroom.
The interaction between syntax and whom demonstrated that the sentences with
subject-extracted clefts and who were rated higher than those with whom; sentences with
object-extracted clefts showed the reverse (whom sentences were rated higher than who
ones). Given that these results show that participants seemed to accept the prescriptive
rules regarding the use of whom, I feel that my teaching of whom and how it is properly
used should not be discontinued. Participants seem to have easily recognized that
sentences like *It was I whom gave the speech were wrong, while sentences such as It is he
whom the lion injured were grammatically acceptable. Some English teachers and other
grammarians have believed that whom will disappear quickly from use; its demise has
been repeatedly foretold in papers written within the past 70 years, including in the
early 21st century (e.g., Miller, 1957; Frank, 1962; Long, 1975; Redfern, 1981; and Soles,
2005). While it may be disappearing, I do not agree that it is disappearing quickly—
while researching this thesis, for example, I encountered many articles from journals in
several academic fields that use whom in the title. For instance, in the PsychINFO
database, there were articles such as “With whom to dine? Ravens’ responses to foodassociated calls …”, “The who and whom of help-giving”, “Whom are you promoting”,
IT-CLEFTS AND THEIR PROCESSING
49
“Who defers to whom and why? Dual pathways linking demographic differences and
dyadic deference to team effectiveness”, and “Alternative diagnoses in patients in
whom the GP considered the diagnosis of pulmonary embolism.” None of these had
anything to do with linguistics or English grammar—the topics included animal
communication, social behavioral psychology, and human resource management—but
their authors recognized the importance of using the accusative whom appropriately.
The frequent use of whom in these and other journal articles demonstrates that it is still a
relevant word in the English language, especially in titles that contain both who and
whom where the duplication of who could be confusing.
I don’t expect my students, or people in general, to incorporate whom into their
everyday conversations, but I think it is important, especially given its still-widespread
use in formal writing, that the rules pertaining to whom should continue to be taught as
part of English grammar. I agree to a certain extent with Derek Soles (2005) when he
says that people tend to label anyone who uses whom in spoken language as “an effete
snob, a pretentious pedant, an English teacher” (p. 34), though I do not agree with his
argument that whom can always be replaced by who: a title such as “The who and who
of help-giving” would make little sense. I want my students to understand the
difference between who and whom, and to know how to use each one appropriately in
academic or professional writing tasks, but I would not expect them to change their
conversational patterns to actively incorporate whom.
The experiment I conducted did answer some questions about case preferences
in clefts, but there are still interesting questions that can be explored. For example, if I
IT-CLEFTS AND THEIR PROCESSING
50
were to run a follow-up experiment in the future, I would add sentences containing that
to the conditions. Given that 16% and 28%, respectively, of the participants used that for
the It was Matilda … and It was Sara … sentence completions, where who or whom were
expected, I believe that sentences containing that in place of who or whom would receive
higher than expected ratings. Plenty of questions still exist to make the continued study
of it-cleft sentences an active topic of research.
IT-CLEFTS AND THEIR PROCESSING
51
References
Aarts, F. (1994). Relative who and whom: Prescriptive rules and linguistic reality.
American Speech, 69(1), 71–79.
Carnie, A. (2013). Syntax: A generative introduction (3rd ed.). West Sussex, UK: John
Wiley & Sons.
Frank, J. F. (1962). Who and whom. College Composition and Communication, 13(4), 35-37.
Geist, R. J. (1952). Professors' English. College Composition and Communication, 3(1), 16-20.
Gordon, P. C., & Lowder, M. W. (2012). Complex sentence processing: A review of
theoretical perspectives on the comprehension of relative clauses. Language and
Linguistics Compass, 6(7), 403-415.
Hartmann, K., & Veenstra, T. (2013). Introduction. In K. Hartmann & T. Veenstra (Eds.),
Cleft Structures (pp. 1–32). Philadelphia: John Benjamins.
Kiss, K. É. (1998). Identificational focus versus information focus. Language, 72(2), 245–
273.
Long, R. B. (1975). Problems in English grammar. TESOL Quarterly, 9(2), 197–202.
MacKay, D. G. (1980). Psychology, prescriptive grammar, and the pronoun problem.
American Psychologist, 35(5), 444–449.
McM., J. B. (1945). Who and Whom. College English 7(2), 104–105.
Miller, C. G. (1957). Is correct English worth the price? The Clearing House 32(3), 135–136.
Patten, A. L. (2012). The English it-cleft: A constructional account and diachronic
investigation. Boston: De Gruyter Mouton.
IT-CLEFTS AND THEIR PROCESSING
52
Redfern, R. K. (1981). The language game: The death of whom? The English Journal 70(4),
82-83.
Reeve, M. (2013). The cleft pronoun and cleft clause in English. In K. Hartmann & T.
Veenstra (Eds.), Cleft Structures (pp. 165–185). Philadelphia: John Benjamins.
Soles, D. (2005). Who needs whom? English Journal 94(5), 34.
Straus, J., Kaufman, L., & Stern, T. (2014). The blue book of grammar and punctuation (11th
ed.). San Francisco: Jossey-Bass.
Warren, T. and Gibson, E. (2005). Effects of NP type in reading cleft sentences in
English. Language and Cognitive Process 20(6), 751–767.
Wolfram, W. and Schilling-Estes, N. (1998). American English: Dialects and variations.
Malden, Massachusetts: Blackwell Publishers, Inc.
IT-CLEFTS AND THEIR PROCESSING
Appendix A
Experiment Sentences
Item 1
a. It was I who completed the assignment.
b. It was I whom completed the assignment.
c. It was me who completed the assignment.
d. It was me whom completed the assignment.
Item 2
a. It was she who painted the portrait.
b. It was she whom painted the portrait.
c. It was her who painted the portrait.
d. It was her whom painted the portrait.
Item 3
a. It was you who broke the lamp.
b. It was you whom broke the lamp.
c. It was you who broke the lamp.
d. It was you whom broke the lamp.
Item 4
a. It is we who won the dance contest.
b. It is we whom won the dance contest.
c. It is us who won the dance contest.
d. It is us whom won the dance contest.
Item 5
a. It is they who bought the bikes.
b. It is they whom bought the bikes.
c. It is them who bought the bikes.
d. It is them whom bought the bikes.
Item 6
a. It is he who writes best-selling novels.
b. It is he whom writes best-selling novels.
c. It is him who writes best-selling novels.
d. It is him whom writes best-selling novels.
53
IT-CLEFTS AND THEIR PROCESSING
Item 7
a. It is you who kicked the ball into the window.
b. It is you whom kicked the ball into the window.
c. It is you who kicked the ball into the window.
d. It is you whom kicked the ball into the window.
Item 8
a. It was I who gave the speech.
b. It was I whom gave the speech.
c. It was me who gave the speech.
d. It was me whom gave the speech.
Item 9
a. It is they who play soccer on the weekends.
b. It is they whom play soccer on the weekends.
c. It is them who play soccer on the weekends.
d. It is them whom play soccer on the weekends.
Item 10
a. It is you who caught the fish.
b. It is you whom caught the fish.
c. It is you who caught the fish.
d. It is you whom caught the fish.
Item 11
a. It was he who recommended the movie.
b. It was he whom recommended the movie.
c. It was him who recommended the movie.
d. It was him whom recommended the movie.
Item 12
a. It was she who figured out the problem.
b. It was she whom figured out the problem.
c. It was her who figured out the problem.
d. It was her whom figured out the problem.
Item 13
a. It was we who the police questioned.
b. It was we whom the police questioned.
c. It was us who the police questioned.
d. It was us whom the police questioned.
54
IT-CLEFTS AND THEIR PROCESSING
Item 14
a. It was they who Sally met at the restaurant.
b. It was they whom Sally met at the restaurant.
c. It was them who Sally met at the restaurant.
d. It was them whom Sally met at the restaurant.
Item 15
a. It is he who the bus driver accused of theft.
b. It is he whom the bus driver accused of theft.
c. It is him who the bus driver accused of theft.
d. It is him whom the bus driver accused of theft.
Item 16
a. It was you who Jerry invited to the party.
b. It was you whom Jerry invited to the party.
c. It was you who Jerry invited to the party.
d. It was you whom Jerry invited to the party.
Item 17
a. It was I who the students voted for.
b. It was I whom the students voted for.
c. It was me who the students voted for.
d. It was me whom the students voted for.
Item 18
a. It was she who the dog followed into the park.
b. It was she whom the dog followed into the park.
c. It was her who the dog followed into the park.
d. It was her whom the dog followed into the park.
Item 19
a. It is you who the cat scratched yesterday.
b. It is you whom the cat scratched yesterday.
c. It is you who the cat scratched yesterday.
d. It is you whom the cat scratched yesterday.
Item 20
a. It was we who the mayor thanked.
b. It was we whom the mayor thanked.
c. It was us who the mayor thanked.
d. It was us whom the mayor thanked.
55
IT-CLEFTS AND THEIR PROCESSING
Item 21
a. It was she who the teacher sent to the principal's office.
b. It was she whom the teacher sent to the principal's office.
c. It was her who the teacher sent to the principal's office.
d. It was her whom the teacher sent to the principal's office.
Item 22
a. It is he who the lion injured.
b. It is he whom the lion injured.
c. It is him who the lion injured.
d. It is him whom the lion injured.
Item 23
a. It was I who the guide answered first.
b. It was I whom the guide answered first.
c. It was me who the guide answered first.
d. It was me whom the guide answered first.
Item 24
a. It was them who the company paid off.
b. It was them whom the company paid off.
c. It was they who the company paid off.
d. It was they whom the company paid off.
56
IT-CLEFTS AND THEIR PROCESSING
57
Appendix B
Minor Patterns
Participants giving ungrammatical fillers high ratings
Forty-five (45) participants rated clearly ungrammatical filler cleft sentences (e.g.
*It was the softball who broke the window.) with either a 6 or a 7; removing their responses
from the ratings of the experiment sentences dropped the average ratings. This was
unexpected; I anticipated that if they rated ungrammatical sentences highly, they might
assign the more natural/acceptable sentences lower ratings, so removing their
responses would raise, not lower, the averages.
Participants misusing who in fill-in clefts
Two participants misused the pronoun who in the fill-in cleft sentence completion
of the It was the car … sentence, writing *It was the car who needed a new engine. None of
the experiment sentences misused who in the same way, since the pronoun it was never
used in the cleft phrase position. Interestingly, even though these participants used who
erroneously with an inanimate object, their average rating for the same sentence
structure (subject/nominative/who)—which involved animates instead of inanimates—in
the experiment sentences was appreciably higher than the average rating of the group
as a whole (4.83 vs. 4.47).
Participant using no relative pronoun
There was one participant who used no pronouns at the beginning of the cleft
clauses (e.g., *It was Matilda for president of the student body and *It was the car needed a new
engine.) This participant gave extremely high ratings for all of the sentences; perhaps the
IT-CLEFTS AND THEIR PROCESSING
poorly written sentence completions reflect a usage misunderstanding of using
who/whom/that in cleft clauses.
58